Frederick Ross Email & Phone Number

Bellevue, Washington, United States

Greater Seattle Area

Redmond, Washington, United States

Greater Seattle Area

Seattle, Wa

Seattle, Wa

I worked on essentially every part of SignalSense's network security product.I worked in the main system, a daemon written in Go. I wrote the initial infrastructure for running models over captured network traffic. I worked on data extraction in the network capture pipeline. I worked on the REST API and the front end.Outside the system, I set up deployment of the product to sensors (TLS secured apt repositories). I built log and metric aggregation and the dashboards to monitor deployed sensors.I also served as the communication layer between the data science team and the engineering team, since I could operate on both sides fluently. I only did a little model development myself since the company needed my time on engineering more, but I introduced a number of the methods that became mathematical mainstays of the team, mentored the data scientists in writing deployable code, and acted as a reviewer and sounding board for that team.I also became frustrated with my lack of understanding of MVC and building graphical interfaces while I was here, and did some focused work to rectify that.

Seattle, Wa

I was one of the early engineers in Splunk's developer tools group. I worked on the Splunk SDK for Python, the Splunk SDK for Java, and the Splunk SDK for C# with very tightly knit teams of one or two other engineers. I built the Splunk SDK for Ruby, the Splunk extension for Visual Studio, and the Splunk plugin for Eclipse by myself, and did the engineering to turn the Splunk library for Java logging and Splunk's JMX from prototypes to releasable products.As a group we also documented the Splunk's REST API, effectively tested it for regressions with our SDK test suites, and I did a lot of reviewing of API changes and drafts for the core product team.My colleagues at Splunk were very encouraging as I learned how businesses work and how my engineering knowledge transferred to teams. I mentored a several junior engineers, who have blossomed, I'm proud to say. I worked booths and conferences, did some evangelism for Splunk in academic settings where I had more background than most of my colleagues, and acted as a resource on statistics and mathematics for people throughout the company.

Seattle, Wa

This position was a poor fit for me. I really liked my colleagues in lab medicine, but I was intellectually isolated and frustrated with the incidental complexity of bioinformatics work. Most of my time was spent automating the construction of building reference sets of sequences from public databases and using them construct species profiles from deep sequencing of microbiota. The primary application was some very interesting work on deep sequencing of vaginal microbiota in vaginosis.At the end of my time here, I also did an evaluation of the (then) new Illumina MiSeq high throughput sequencer, which had a radically different error profile from the sequencers most alignment algorithms were designed for. I managed to improvise some workable algorithms for that error profile from existing primitives.

Lausanne, Switzerland

I left grad school with a masters degree, and moved to a staff scientist role in one of EPFL's core facilities where I acted as a consultant on statistics and bioinformatics for scientists in the department and wrote software to improve the core facility's workflows. During this time I was also asked to write some columns on statistics for clinicians by the Journal of Psychiatric Practice.When I arrived, the facility's workflows were a patchwork of R and bash scripts, and the workflow consisted of tweaking these scripts, rerunning the analysis, perhaps tweaking a downstream script and running it over the output of the first, and proceeding to drown in input and output files. The facility's director, Jacques Rougemont, asked me to improve that.We first tried a centralized, web based system. I made a prototype in Erlang, but found its poor integration with the Unix world around it prohibitive. I rewrote the prototype in Haskell, and discovered that the cluster bioinformaticists used in the Lausanne area had a version of libc over a decade old, too old to compile the Haskell compiler. We learned a lot from those initial iterations, though. The next version was not centralized. It was run by individual analysts on their own machines. It tracked individual executions and the files used and created by them, provided a simple file management system, and could dispatch arbitrary code across the cluster.Then the Swiss government decided that my visa was coming to an end, two years earlier than any of us had thought. The system was written in PLT Racket, and I was the only Scheme programmer in the group. While I was there we had converged strongly on Python, though. I sat down one Sunday afternoon on my couch and wrote the most useful core functionality in about 500 lines of Python. We elaborated it over the next couple of months but that code hasn't changed since (probably the only system I can say that about). The system is released under the name of 'bein'.

New York, Ny

In July 2007, the laboratory relocated to the École Polytechnique Fédérale de Lausanne in Switzerland, where I continued, though I was still on the books as a fellow at Rockefeller.I studied the response of Mycobacteria, the genus that includes the etiological agent of tuberculosis, to antibiotics. In particular, I studied subpopulations that did not die in the presence of the drug, but whose offspring did not inherit any resistance to it, which is one of the major suspected causes of the long treatment time that tuberculosis requires.Along the way I became a mediocre molecular biologist and a passable microbiologist, was qualified to work in Biohazard Level 3 facilities, reverse engineered and wrote control software for automated microscopes, developed some skill at computer vision, and taught myself a lot of statistics (to the point where I became the go-to guy in the department on the subject). I built a bunch of random systems for the laboratory, such as a web app to capture and display temperature from our bacterial incubator, and a program to generate grids on microscope stages for experiments.I also did a lot of other random things. I took graduate electromagnetism at Columbia. I studied thermodynamics, something my undergraduate degree had basically skipped, on the side. I spent a lot of time reading Dijkstra's work on calculational methods. I wrote a lot of Haskell and was quite active in that community where I was mostly known for doing things like serial port control and image capture from high end CCDs as part of my microscopy work.This is also the period when I started transitioning from a programmer to a software engineer. I read Fred Brooks's 'The Mythical Man Month' and everything else I could lay my hands on. I learned to use git. I started doing extensive, automated testing. I learned to write requirements and specifications. I planned and estimated projects, and measured my execution to improve my estimation.

Charlottesville, Va

Ran problem sessions, prepared solutions, and graded homework for introductory physics for honors engineering students under Keith Williams and widely applied physics course for third year physics majors under Eugene Kolomeisky.

Charlottesville, Va

I worked in this research lab part time while studying physics and math. I wrote a lot of miscellaneous code for them. For example:The group had 120TB of data stored in a tape library at NERSC from the HyperCP experiment. They needed subsets a few hundred megabytes in size extracted from it for the doctoral students and postdocs. I was assigned to program and run that extraction using the tapes and Linux worker nodes at NERSC, an HP-UX box in Berkeley where I could run PostgreSQL, and a Digital Unix server in Charlottesville.Today we would call the solution map-reduce. In practice it was a master Perl program---Perl, the only language I had access to on all the systems involved---on the local DEC box that tracked the available slots in the tape library and dispatched workers, also written in Perl, on the Linux cluster at NERSC to process the tapes and write the extracted subsets onto durable storage.At another time, the group needed a way to test detectors and their attached electronics as they were being assembled. This time I had C, a semi-documented library for interfacing with the Jorway CAMAC controllers they were using (it took three days before I found the footnote that explained how to make that particular controller work), and a Linux machine. I wrote them a nice shell that would turn things on and off, and poll data off the system to the screen and to a file.I also learned a lot about high rate data acquisition and designing systems to keep up with high speed flow. That keeps coming in handy.

Independence, Va

LSNet gave me my formative education in computers. I was a jack of all trades here. I built networks. I administered Linux systems. I learned C and added database backend support to a RADIUS server. I wrote one of the first web pages that let customers configure a computer and see what the price would be.My boss, Jim Tarvid, had a long history in computers, dating back to working as a service engineer on Burroughs large systems. He encouraged me to wander and occasionally directed me. I was lucky enough to work through such wonderful but out of the way books as Ullman's 'Elements of Standard ML' and Brodie's 'Learning Forth' and 'Thinking Forth'.

Msc, Biology

advisor: John McKinney

Bsc/Ba, Physics And Mathematics

Degrees awarded with highest distinction; mathematics degree was mostly graduate coursework Echols scholar πµε mathematics honors society.